extracting mapped reads from the bam file

Question

[KVC_bioinfo]

Hello all,

I am trying to extract the mapped reads from a bam file for a specific region.

samtools view -h in.bam "chr1:regionstart-regionend" > out.bam

this gives me all the reads falling into that region. (which contains read that mapped into that region plus reads unmapped into that region)

I am looking for reads that mapped to that specific region. Which options can I include?

Or samtools in unable to do it?

Thank you in advance.

[swbarnes2]

How are you distinguishing between reads that "fall" in a position, from reads that "map" to it?

[KVC_bioinfo]

That is basically I am not able to do it. I am interested in extracting reads mapped in a specific region.

[cpak1981]

To most of us, there is no difference in what you've done and what you're asking for, so clarification is required. Please elaborate on what the difference is between what you've done and what you're asking for by editing your original post.

[KVC_bioinfo]

Thank you. I added more explanation.

[Pierre Lindenbaum]

please, stop asking questions and acknowledge the answers (green tick on the left) or comment the answers:

• Understanding SAM file format
• Loading genome on IGV
• Extracting reads from given region
• GRCh37/hg19 human genome sequences from NCBI
• Alignment in IGV: extracting specific reads.
• extracting reads from sam/bam file
• (...)

[KVC_bioinfo]

Hello Pierre,

All the posts you pulled here does not answer the question I asked in this post. I have always acknowledged the answer.

[cpak1981]

Perhaps you're unaware of how to acknowledge answers. There should be a check mark (in addition to up/down arrows) that you as the OP can toggle on/off.

[Pierre Lindenbaum]

no: eg.: Understanding SAM file format answer is correct but not commented or not flagged as answered.

[KVC_bioinfo]

Got it. Changed it to answer. Thank you.

Answer 1

[Pierre Lindenbaum]

using http://lindenb.github.io/jvarkit/VcfFilterJdk.html after samtools to get the reads strictly in the regions:

samtools view -b -F 4 in.bam "chr1:regionstart-regionend" |\
java -jar dist/vcffilterjdk -e 'return record.getStart()>=regionstart &&  record.getEnd()<=regionend;'

[KVC_bioinfo]

This works well. Thank you very much.

[KVC_bioinfo]

Thank you very much. I will try this out.

Answer 2

[Devon Ryan]

samtools view -b -F 4 in.bam "chr1:regionstart-regionend" > out.bam

I assume by "unmapped in a region" you mean "having coordinates in a region but also being unmapped". So, just filter unmapped reads.

[KVC_bioinfo]

Suppose I have a long read which maps to few exons of a gene A and it does not map to rest of the exons of gene A. Now when I extract the reads that map to say exon 1, I want to have only reads that mapped to a region of exon one. When I use the command you gave, that extracts the reads falling in that regions. however, few reads in that region do not map to the region specified but they map somewhere else on the gene.

I could see that on IGV few genes were not mapping to the region I specified but still present in the bam file I got from above command.

[Pierre Lindenbaum]

few reads in that region do not map to the region specified but they map somewhere else on the gene.

it's not clear if you're talking about a read or about a (paired) read and its' mate.

[KVC_bioinfo]

my original reads are single end reads coming from nanopore sequencing.